Adaptive strategy for superpixel-based region-growing image segmentation

Chaibou, Mahaman Sani; Conze, Pierre-Henri; Kalti, Karim; Solaiman, Basel; Mahjoub, Mohamed Ali

doi:10.1117/1.jei.26.6.061605

Cited by 21 publications

(6 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To address the problem of over-or undersegmentation that occurs when SLIC is used to segment areas of vastly different sizes, hierarchical clustering is applied. The benefits of clustering superpixels into larger regions were shown in [8]. In this work, superpixels are described by the mean value of their pixels, which can either be a color or a feature vector.…”

Section: Hierarchical Clusteringmentioning

confidence: 99%

Superpixel Pre-Segmentation of HER2 Slides for Efficient Annotation

Öttl¹,

Mönius²,

Marzahl³

et al. 2022

Preprint

View full text Add to dashboard Cite

Supervised deep learning has shown state-of-the-art performance for medical image segmentation across different applications, including histopathology and cancer research; however, the manual annotation of such data is extremely laborious. In this work, we explore the use of superpixel approaches to compute a pre-segmentation of HER2 stained images for breast cancer diagnosis that facilitates faster manual annotation and correction in a second step. Four methods are compared: Standard Simple Linear Iterative Clustering (SLIC) as a baseline, a domain adapted SLIC, and superpixels based on feature embeddings of a pretrained ResNet-50 and a denoising autoencoder. To tackle oversegmentation, we propose to hierarchically merge superpixels, based on their content in the respective feature space. When evaluating the approaches on fully manually annotated images, we observe that the autoencoder-based superpixels achieve a 23% increase in boundary F1 score compared to the baseline SLIC superpixels. Furthermore, the boundary F1 score increases by 73% when hierarchical clustering is applied on the adapted SLIC and the autoencoder-based superpixels. These evaluations show encouraging first results for a pre-segmentation for efficient manual refinement without the need for an initial set of annotated training data.

show abstract

Section: Hierarchical Clusteringmentioning

confidence: 99%

Superpixel Pre-Segmentation of HER2 Slides for Efficient Annotation

Öttl¹,

Mönius²,

Marzahl³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Each lime fruit's edges are detected using the Canny algorithm as shown in Figure 7, [25], [26].  Step 3: Image segmentation using separate regions of each lime fruit and background with a regiongrowing method as shown in Figure 8, [27], [28].…”

Section: Learning Phasementioning

confidence: 99%

Image-based lime size grading using the comparison ratio of the pixel radius and the actual size of lime fruit

Chimlek

Jitanan²

2021

IJEECS

View full text Add to dashboard Cite

Lime is a commercially important fruit in Thailand whose sale price depends on the fruit’s size; hence, farmers must grade limes by size before distribution. However, as lime grading machines are very expensive and each province has different size grading limits, grading is often performed manually, which is time-consuming and error-prone. Agricultural production systems for automatic selection and grading use image processing techniques for extracting key features. Therefore, this study proposes techniques to extract features of limes and to develop analytical methods for grading them. This method can reduce time and cost, and increase accuracy and flexibility for selecting different lime sizes according to each province’s size criteria. To verify our method, we classified limes according to criteria from four Thailand provinces as sample data in an experiment. The focal image feature was the radius or diameter of the lime and the grading conditions were defined by the maximum comparison ratio of the fruit’s radius in pixels to the measured radius of the actual lime in centimeters. The average grading accuracy was 99.59%, which outperformed that of mechanical grading. The processing time was 1.70 seconds per individual fruit.

show abstract

“…A feature vector was calculated for each superpixel which comprises 9 intensity, texture and gradient based statistical features (mean, variance, skewness, 10-bin intensity histogram, contrast, energy, entropy, 10-bin histogram of gradient orientation and 10-bin histogram of gradient magnitude). Inspired by Chaibou et al (2018) [16], the proposed merging procedure is based on the agglomerative clustering algorithm via the Ward method [20]. The superpixel similarity measure is defined in the same way as Chaibou et al (2018), where both content similarity and border similarity make up the overall similarity measure.…”

Section: Pseudo-label Refinementmentioning

confidence: 99%

“…However, in the case of pathology, such as brain tumour segmentation, a circular shape prior is not suitable as there is a large variance in topology and extent of disease between patients. Chaibou et al (2018) [16] developed a strategy for unsupervised segmentation of natural images based on superpixel clustering. Superpixels to be merged are assumed to satisfy two important criteria: spatial adjacency and perceptual similarity.…”

Section: Introductionmentioning

confidence: 99%

Pseudo-label refinement using superpixels for semi-supervised brain tumour segmentation

H.¹,

Caterina²,

Voisey³

2021

Preprint

View full text Add to dashboard Cite

Training neural networks using limited annotations is an important problem in the medical domain. Deep Neural Networks (DNNs) typically require large, annotated datasets to achieve acceptable performance which, in the medical domain, are especially difficult to obtain as they require significant time from expert radiologists. Semi-supervised learning aims to overcome this problem by learning segmentations with very little annotated data, whilst exploiting large amounts of unlabelled data. However, the best-known technique, which utilises inferred pseudo-labels, is vulnerable to inaccurate pseudo-labels degrading the performance. We propose a framework based on superpixels -meaningful clusters of adjacent pixels -to improve the accuracy of the pseudo labels and address this issue. Our framework combines superpixels with semi-supervised learning, refining the pseudo-labels during training using the features and edges of the superpixel maps. This method is evaluated on a multimodal magnetic resonance imaging (MRI) dataset for the task of brain tumour region segmentation. Our method demonstrates improved performance over the standard semi-supervised pseudolabelling baseline when there is a reduced annotator burden and only 5 annotated patients are available. We report DSC=0.824 and DSC=0.707 for the test set whole tumour and tumour core regions respectively.

show abstract

Adaptive strategy for superpixel-based region-growing image segmentation

Cited by 21 publications

References 36 publications

Superpixel Pre-Segmentation of HER2 Slides for Efficient Annotation

Superpixel Pre-Segmentation of HER2 Slides for Efficient Annotation

Image-based lime size grading using the comparison ratio of the pixel radius and the actual size of lime fruit

Pseudo-label refinement using superpixels for semi-supervised brain tumour segmentation

Contact Info

Product

Resources

About